Requesting customize_embedding api failed without any error message

[ZZfive]

Using the token obtained through email, I can successfully execute "python demo_examples/generic_inference.py --token ", but "python demo_examples/customize_embedding.py --token " always fails. The error message is as follows, and there is no error message.

[Mountchicken]

Hi @ZZfive I just updated the code and please pull the newest code. It should work fine now.

[ZZfive]

Hi @ZZfive I just updated the code and please pull the newest code. It should work fine now.

Thanks for your response! After I pull latest code, I successfully executed "python demo_examples/customize_embedding.py --token ", return the url of the embedding model

[ZZfive]

Hi @Mountchicken The "python demo_examples/customize_embedding.py --token " mentioned above was executed successfully, Only the token was set without any modification. However, when the prompt_image set in prompts was set to a local image, the request still failed, the output is as follows:

When the above failed, 8 pictures were set in prompts, and then the pictures were reduced to 4. The request was successful and the url of the embedding model was returned. Are there any specific restrictions on the number and size of images when requesting customize_embedding?

[Mountchicken]

The maximum number of images for customize_embedding API is 4, and we do not restrict the image size.

[ZZfive]

The maximum number of images for customize_embedding API is 4, and we do not restrict the image size.

Is the maximum number of images for all API 4, such as generic_inference and interactve_inference?

[Mountchicken]

Yes